Earlier this year, Major League Baseball reached an agreement with the Atlantic League that allowed MLB to test any new playing rules in Atlantic League games. Five new rules already went into effect in the first half of the season, and a day after the Atlantic League All-Star Game, in addition to testing the electronic strike zone using TrackMan radar, MLB announced four new rules that would go into effect in the second half of the Atlantic League season. One of the four seems to have received stronger criticism on the Internet than others, and that is the idea of stealing first base.
According to this new rule, if a pitch is not caught in flight, the batter has the right to “steal” first base. In essence, this is an extended version of the dropped-third-strike rule that has long been engrained in the baseball rule books. Of course, a dropped third strike has many criteria: it must occur with two strikes, the pitch must be a strike, and unless it occurs with two outs, there cannot be a runner occupying first base. With this new rule, however, batters now have the right to go for first base in any count, regardless the pitch is a ball or strike, and in any base situations. And recently, we witnessed the first steal of first in professional baseball history:
— SoMD Blue Crabs (@BlueCrabs) July 14, 2019
On the second pitch in the bottom of the sixth inning, Alejandro Chacin of the Lancaster Barnstormers threw a wild pitch, which allowed Southern Maryland Blue Crabs outfielder Tony Thomas to steal first base. From the video, we can clearly tell that the players were still adjusting to this new rule, as neither the catcher nor the batter reacted at first. The fact that catcher Anderson De La Rosa took time to react was probably the main reason that allowed Thomas to reach first without a throw. Interestingly, the term “stealing first” might actually be a misnomer, as the play was scored as a fielder’s choice and counted as an 0-for-1, according to Somerset Patriots southpaw Rick Teasley.
We were told it counts as an 0 for 1, fielder’s choice.
— Rick Teasley 銳力 (@ItAintEasy37) July 14, 2019
For now, this new rule has not occurred too many times in the Atlantic League yet, so we don’t really have enough data to be able to analyze its effect. However, I wanted to utilize the dropped third strike rule that has long been in effect, so I used MLB data from the past four seasons to study how this new rule might affect gameplay. From the Retrosheet play-by-play log, we can identify every plate appearance that ended in a strikeout (EVENT_CD == 3), and those that resulted in the batter either reaching base (BAT_DEST_ID != 0), or putout at first (BAT_PLAY_TX != ‘NA’ & BAT_PLAY_TX != 2).
|Season||Ran Base||Out at First||Safe at First||Safe Rate||MLB SB%|
In the past four seasons, when a third strike is dropped, major league hitters took off for first base slightly more than 1,000 times each year, but they only had a roughly 11% chance of successfully reaching base. Compared to the stolen-base success rate, which hovers around the 70-73% range, reaching first base on a dropped third strike is obviously much harder, and rightfully so. But in addition to the safe rate when the batter does go for first, I was also curious, when given the opportunity, how often do batters actually run towards first base, instead of just being tagged out by the catcher?
From the Retrosheet play-by-play log, there’s no way for us to identify which strikeouts were dropped and the batters were immediately tagged out. Therefore, I went on Baseball Savant’s Search tool, set the filters to match conditions in which dropped third strikes are possible, and downloaded every two-strike pitch that was labeled “Swinging Strike (Blocked).” Theoretically, these are all the pitches in which the batter has the right to go for first base, and if the description of the play is less than 50 characters long, it means the batter was tagged out by the catcher and did not require any throws.
However, this selection method neglected some pitches that were labeled as simply a “Swinging Strike” because the catcher was not able to block it at all, and some called strikes that the catcher did not catch in flight. I believe none of these resulted in the batter being tagged out, but some did result in either the batter reaching base or at least required the catcher to throw to first for the putout. With that in mind, I added these pitches (description > 80 characters) to the “Swinging Strike (Blocked)” dataset mentioned above, just so we can have a more accurate denominator when making calculations.
|Season||Opportunities||Ran Base||Run Rate||Reached||Reach Rate|
Overall, when a third strike is dropped, MLB hitters run towards first base around 27-28% of the time, which also means that more than 70% of the time, the batter is tagged out at the plate. The previous calculations showed that when the batter does run, he is safe about 11% of the time, but when we consider all occasions where the batter is technically allowed to run, we only see a reach rate of roughly 3%. Moreover, these are situations where the batter is out anyway if he doesn’t run, so when we apply this to a non-two-strike situation, the story becomes much different.
The biggest difference with this new rule and the existing dropped-third-strike rule is that the batter has a choice. When the pitch is not caught in flight, the batter has the right to run to first, but is he chooses not to, he can still continue to bat. On the other hand, if he does attempt to go for first, there is still a chance that he would be thrown out. The batter would essentially have to be willing to give up his chance of hitting and try to get on base with their legs, which honestly isn’t much different than bunting for a hit, and that’s already very rare.
For players that wouldn’t even attempt to steal second base, there’s little reason for them to suddenly be willing to “steal first,” which may be even harder to do. Also, as mentioned, this rule is scored as a fielder’s choice and 0-for-1, which means that a hitter’s on-base percentage would decrease, even if he successfully reaches base. It might be very difficult to convince them to try for it, especially for great hitters who do not have good speed, unless they are very certain that they can reach and it would greatly benefit their team.
For players who do not hit very well and have good speed, it may be worth trying, but the role of speed in this play may not be as important as you would think. When bucketing the different strikeout results, I also joined the tables with the Sprint Speed database from Statcast to see if players with better speed reach base more often. Interestingly, the results show that when batters run for first on a dropped third strike, their speed has very little to do with whether they reach base. Perhaps how far the pitch gets away is a much bigger factor.
|Season||Tagged Out||Ran Base||Out at First||Reached Base|
At the same time, I also noticed that players who do run towards first have a higher average sprint speed than players who do not. Even though the 0.3-0.4 feet per second difference seems small, the fact that we have observed a similar gap in all four seasons might suggest that this may be statistically significant. And it makes sense when you think about it, as players who are fast generally have better baserunning awareness, thus maybe can react quicker than players with average or slower speed.
If this new rule were to go into effect in MLB, the group of players that it would benefit most would be those that are poor hitters but have elite speed, such as Billy Hamilton or Dee Gordon. Since they are already bad at getting on base, they can take advantage of this rule and avoid making outs when hitting, and then proceed to wreak havoc on the basepaths. In today’s baseball, these types of players that survive relying on only defensive and baserunning abilities are becoming increasingly rare, so if this new rule can boost the chances of putting them on base and creating more action, it may not necessarily be a bad idea at all.
Note: If anyone would like to view the simple codes I wrote for this project and access the CSV files used, please visit my personal GitHub. All files are stored in the “steal_first” repository.